Surrogate Measures of Operating Cash Flows and Accruals : Problems and Solutions

We examine the measurement error in surrogate measures of operating cash flows and accruals widely used in the accounting literature. We decompose the measurement error into two components, non-articulation error and omission error. We find that the measurement error is large in magnitude and primarily driven by the omission error. Furthermore, we find non-articulation events have little effect on the omission error. We demonstrate that the measurement error produces biased coefficients and incorrect inferences in four common research settings. We also provide a simple correction that eliminates the omission error and show that the correction reduces biases and improves inferences. Importantly, our decomposition and correction can be used by future researchers to identify, understand, and take steps to mitigate the measurement error and the resulting biases that are introduced when using surrogate measures of operating cash flows or accruals